PhD candidates and subjects (December 2020)
The following data are published in accordance to paragraph 4, article 39, of regulation 4485/2017:
The names of the PhD candidates and the members of the advisory committees, are published together with the titles and abstracts of the dissertations, on the Institution’s website, both in Greek and English.
PhD Candidate: Vassilios D. Vassios
Supervisor:
Dimitrios K. Papakostas, Professor IHU
Advisory Committee:
Alkiviadis Hatzopoulos, Professor AUTH
Argirios Hatzopoulos, Assistant Professor IHU
Title:
Fault Testing Methods and Algorithms for Analog and Mixed-Mode Electronic Circuits using Embedded Systems
Summary:
The scope of the PhD Thesis is to develop novel methods and algorithms for testing faults in analog and mixed-mode electronic circuits. The primary methodology that will be followed, consists of data gathered from the measurements of signal response of non-faulty electronic circuits and compared against similar measurements from circuits that are injected with selected faults. The aim of this research is to try to develop new metrics that will help the improve the efficiency of the developed algorithms and try to improve previous algorithms regarding the classification of the Circuit Under Test (CUT) to Good/Faulty by decreasing the percentage of the classified CUTs to False Positive (Faulty) or False Negative (Good). This algorithm will be implemented in an embedded system (μCU,FPGA) with main goal to improve the classification timing of the CUT’s to Good/Faulty.
PhD Candidate: Dimitrios K. Gerontitis
Supervisor:
Panagiotis Tzekis, Associate Professor IHU
Advisory Committee:
Efstathios Antoniou, Associate Professor IHU
Nicholaos Karampetakis, Full Professor AUTHU
Title:
Study of Time Invariant and Varying Systems Using Recurrent Neural Networks
Summary:
The scope of the PhD Thesis is to develop faster families of recurrent neural networks to accelerate the speed of convergence time over previous models. To be precise, we will optimize the existing functions in order to achieve faster convergence speed in finite time, to solve time varying and invariant problems such as, computation inverse of a matrix, solution of important matrix equations such as: Sylvester, Lyapunov, finding generalized inverse, solving tensor equations via Einstein product, solving non-linear equations and etc. The results and conclusions which will emerge from this PhD Thesis are expected to be particularly useful in applications, which are found in areas, such as: Robotics, Image Processing and Optimization Problems. Computer code will be developed in the Matlab’s environment to validate results and simulate specific applications.
PhD Candidate: Georgios Gravanis
Supervisor:
Konstantinos Diamantaras, Professor IHU
Advisory Committee:
Simira Papadopoulou, Professor IHU
Michail Salampasis, Professor IHU
Title:
Fault Detection in industrial / production processes with Deep Learning methods
Summary:
The scope of this Ph.D. thesis is to examine the possibility of using deep neural networks with architectures, such as those of Time Delay Neural Networks (TDNN) and Long Short-Term Memory (LSTM) for early fault detection in industrial and production processes. With this thesis, deep learning architectures that will make the produced models capable of use in real production time, will be developed. Also, the use of appropriate metrics for proper result evaluation, as well as feature extraction methods for effective network training, will be investigated. Finally, unsupervised and reinforcement learning algorithms will be utilized for providing solutions in real-life large-scale applications.
PhD Candidate: Delimaras Vasileios
Supervisor:
Spasos Michalis, Associate Professor IHU
Advisory Committee:
Papakostas Dmitrios, Professor IHU
Hatzopoulos Argirios, Assistant Professor IHU
Title:
Real Time Measurements of Oils Quality, based on the Changes in Electrical Characteristics
Summary:
The scope of the PhD Thesis is to develop a novel method to measure capacitance, suitable for Interdigital Capacitance sensors (IDC). To begin with, an IDC sensor will be designed, that will be able to estimate the quality of an edible or lubricating oil or/and the degree of degradation, due to its continuous use in conditions that cause the degradation. The degradation of an oil is directly related to the changes in electric permittivity. As part of this research, the design of an IDC sensor will be implemented, so that changes in the electric permittivity of an oil are converted into changes in capacitance. Regarding the design of the sensor, the physical properties of the electrode material will be studied, as well as the substrate, the dimensions of the structure, the design geometry (parallel coplanar electrodes, fractal, or other more complex structure) and the degrees of freedom that may exist, in order to increase the sensitivity of the sensor. Furthermore, a novel capacitance measurement technique will be developed, using one-shot methods and capacitance multiplier techniques, to implement a Capacitance-to-Voltage converter (C-to-V Converter). The main goal of this thesis is for a device to be created, that will estimate the quality of an oil via the changes of its electrical properties in real time.
PhD Candidate: Marina B. Delianidi
Supervisor:
Konstantinos I. Diamantaras, Professor IHU
Advisory Committee:
Evangelidis Georgios, Professor University of Macedonia
Sidiropoulos Antonios, Assistant Professor IHU
Title:
Predicting Student Performance Using Time-Dependent Machine Learning Methods and Recommending Educational Content
Summary:
In the field of education and especially in e-learning, through the huge volume of the educational information disseminated on the World Wide Web, the correct recommendation of both a series of courses and educational materials is valuable information for the evolution of students’ educational level. The prediction of students’ knowledge state is the most important information for the successful recommendations of educational content that will contribute to both the improvement and the progress of knowledge state. The aim of the doctoral dissertation is the research of Machine Learning methods for the dynamic assessment of student performance and the development of Recommendation Systems for recommendation educational content to the positive progress of the students’ knowledge state.
PhD Candidate: Pantelis I. Kaplanoglou
Supervisor:
Konstantinos Diamantaras, Professor IHU
Advisory Committee:
George A. Papakostas, Professor IHU
Ignatios Deligiannis, Professor IHU
Title:
Explainable Machine Learning for Intelligent Systems
Summary:
A crucial issue towards widespread application of Machine Learning models is the capability of explaining their functionality and the causes that drive their decisions. The new research area of Explainable Machine Learning offers methods that produce evidence of the system’s behavior in understandable means for humans. By supplying visualizations, metrics and mathematical tools, the understanding on the general functionality of a model, which is based on the formal definition of the method, is expanded to an analytic explanation of its internal characteristics. The non-explainable Deep Neural Networks show increased accuracy compared to explainable models, but with the downside of functioning like black-boxes, to which we provide an input in order to generate an output. Thoughts about widespread use are accompanied with questions regarding reliability, bias and concerns about ethics, physical security. Additionally, there is a lack of trust in them especially in healthcare, pharmaceutical and biomedical sectors. Potential social ramifications have led legislators to establish the “right to explanation” a provision that affects the applicability of state-of-the-art models in products. In parallel, to implement innovative intelligent systems, the experimental implementations of Machine Learning methods need to evolve into software architectures, taking under consideration additional aspects that concern the new sector of Machine Learning Engineering. Research as part of this doctoral thesis focuses on explanatory methods for existing models and attempts to introduce new explainable models, that will be capable to produce consistent predictive results, that humans both anticipate and understand. Secondarily, will produce new standards for machine learning software that provide explanations.
PhD Candidate: Alkiviadis K. Katsalis
Supervisor:
Konstantinos Diamantaras, Professor IHU
Advisory Committee:
Athena Vakali, Professor AUTH
Michail Salampasis, Professor IHU
Title:
Machine Learning Methods for Natural Language Processing (NLP) and Natural Language Generation (NLG)
Summary:
The scope of the PhD Thesis is the contribution with new ideas and improvements in the area of Natural Language Processing (NLP) and Natural Language Generation (NLG). More specifically, the research will be focused on Automatic Text Summarization. One idea is to enrich word/sentence embeddings in terms of syntax, semantics and the context of the text. The use of Knowledge Graphs will be examined on how useful can be in order to achieve the aforementioned aims. Furthermore, except from the English language, the final implementation is aiming to include approaches for the Greek language as well, where the needs of research in the area of NLP are greater and equally important.
PhD Candidate: Grigorios S. Katsios
Supervisor:
Efstathios Antoniou, Associate Professor IHU
Advisory Committee:
Pangiotis Tzekis, Associate Professor IHU
Stavros Vologiannidis, Assistant Professor IHU
Title:
Cooperative Control and Study of the Structural Properties of Singular Multi-Agent Dynamical Systems
Summary:
The present PhD research programme aims to study the structural properties of multi-agent systems, whose components (agents) are themselves singular systems of first, second or higher order. Singular systems arise naturally in the study of dynamical processes which are either subject to algebraic constraints or in cases where the system itself results from the interconnection of smaller components. The aim of the PhD research is to develop results, both at theoretical and application level, by investigating the structural properties of such systems and by extending existing collaborative control techniques in the class of singular cooperative systems.
PhD candidate: Konstantinos Kelesidis
Supervisor:
Dimitrios Dervos, Professor, IHU
PhD Committee members:
Ioannis Marmorkos, Professor, IHU
Antonios Sidirpoulos, Professor, IHU
Title:
Exploratory and Analytical Processing of Academic Student Records
Summary:
Quality Assurance comprises a necessity for the University academic unit. It is achieved with systematic assessment procedures. The latter involve the use of quantitative and qualitative indicators.
Considering the above, the present PhD research aims to utilize quality indicators that relate to student assessments in courses of the IHU IEE department’s programmes of study. Indicators can be utilized in conducting exploratory data analysis as well for defining measures that in turn can be utilized in the data mining processing of academic records.
An example of a quality measure is the definition and the quantification of a Course Degree of Difficulty (CDoD). Once defined, the latter can be incorporated into the logic of a data mining algorithm for improved course recommendations.
The research is conducted in collaboration with the members of the departmental Internal Evaluation Group (IEG). This is done in order to achieve the highest possible degree of convergence and harmonization between the research aims and the needs dictated by the internal evaluation procedures at a typical university academic unit.
PhD Candidate: Kostopoulos Evangelos
Supervisor:
Diamantaras Konstantinos, Professor, IHU
Advisory Committee:
Ioanna Chouvarda, Professor AUTH
Konstantinos Goulianas, Professor IHU
Title:
Neural network pruning methods and their application on medical imaging
Summary:
For the past few years there have been several attempts to use convolutional neural networks so as to analyze and process medical images (CT, MRI, X-RAY) with very promising results. However, due to missing or incorrect information, improving and restoring images using filtering and finding morphological patterns can be very hard. This PhD will research the application of pruning methods in terms of medical images’ analysis and focus on the design of a new pruning technique. This technique will simulate the human optic nerve, which isolates and recognizes objects using pruning and repeating methods during the entire recognition process. The goal is to improve the performance and reliability of this predictive process, as well as reduce the margin of error in terms of the prediction process that relates to medical image analysis.
PhD Candidate: Lampropoulos Georgios
Supervisor:
Keramopoulos Euclid, Associate Professor, IHU
Advisory Committee:
Diamantaras Konstantinos, Professor, IHU
Evangelidis Georgios, Professor, UoM
Title:
The role of educational technology and gamification in improving education, cognitive and social-emotional development and 21st century skills cultivation: Development and evaluation of virtual and augmented reality applications, artificial intelligence tools and serious games.
Summary:
Nowadays, the rapid technological advances and the digitalization of everyday life have created new educational needs and requirements. The aim of this study is to scrutinize the role of educational technology and gamification in the context of the constantly developing 21st century education and pedagogy. More specifically, this study will examine the way in which the use of emerging technological applications can enrich the contemporary educational process, reinforce wellbeing, promote cognitive and social-emotional development and improve 21st century skills cultivation. Furthermore, it will analyze and present the technologies of augmented reality, virtual reality, artificial intelligence and serious games as well as the contemporary educational approaches of gamification and game-based learning. Through the use of these technologies and approaches and in collaboration with educational communities and institutes intelligent, student-centered and personalized virtual learning environments and applications will be developed and evaluated. Finally, in order to infer valid and reliable conclusions, qualitative and quantitative studies, which will be based on the creation and use of questionnaires and big data analysis, will be included.
PhD Candidate: Konstantinos M. Melisidis
Supervisor:
Euclid Keramopoulos, Associate Professor IHU
Advisory Committee:
Panagiotis Adamidis, Professor IHU
Apostolos Ampatzoglou, Assistant Professor at University of Macedonia
Title:
An educational approach to teaching programming and computational thinking through augmented reality and gamification
Summary:
The scope of the PhD Thesis is the creation of an educationally oriented augmented reality platform, with gamification mechanisms, for teaching pro-gramming and enhancing computational thinking. More specifically, an augmented reality application platform will be implemented, appropriately de-signed according to pedagogical methods, enriched with gamification mech-anisms and technologically oriented to mobile devices for its adoption in a learning environment. Harmonization with modern pedagogical methods will be essential, as will the use of technologies such as AR-LEs, AR-GBL, PBL, and so on. The final goal of the PhD thesis implementation is expected to be particularly useful for the acquisition of computational thinking as a necessary 21st-century skill, as well as for learning programming.
PhD Candidate: Ilias-Nektarios Seitanidis
Supervisor:
Athanasios Iossifides, Associate Professor IHU
Advisory Committee:
Periklis Chatzimisios, Professor IHU
Melpomeni Ioannidou, Associate Professor IHU
Title:
Small cell radio resource management techniques for Internet of Things services improvement in 5G mobile networks.
Summary:
This PhD thesis will study PHY and MAC layer techniques to enhance the Internet of Things services that have differentiated Quality of Service (QoS) requirements in 5G mobile networks. Towards this end, 5G small cell deployments together with novel radio resource management and multiple access techniques will be the main tools to achieve proper network slicing over the air interface resources. The combined management of the radio resources of small cells and normal or macro cells re-enforced by non-orthogonal multiple access (NOMA) techniques will be explored to effectively support multiple data streams of different requirements (e.g. eMBB, URLLC and mMTC use cases) while leaving standard cellular traffic unaffected as much as possible. Spectrum efficiency and energy efficiency will be considered as the main (among others) key performance indicators of the evaluation of the proposed techniques.
PhD Candidate: Vasileios Stamatis
Supervisor:
Michalis Salampasis, Professor IHU
Advisory Committee:
Konstantinos Diamantaras, Professor IHU
Allan Hanbury, Professor TU Wien
Title:
Applied Intelligence for Federated Patent Search
Summary:
The purpose of this PhD is to contribute in the Federated Patent Search field. The research will start with the results merging sub-processes during distributed patent search in which methods for improving the effectiveness and the efficiency will be evaluated. Existing methods will be optimized and new methods will be discovered like machine learning algorithms for optimizing the results. Furthermore intelligent methods will be implemented in the source selection sub process during distributed patent search. Additionally query expansion is another problem which will be examined and new methods will be created using IPC codes for improving the efficiency of the retrieval. Eventually, the new methods will be embedded in a new system for federated patent search and the system will be tested with real users using user studies.
PhD Candidate: Sokratis A. Tselegkaridis
Supervisor:
Theodosios Sapounidis, Assistant Professor IHU
Advisory Committee:
Aristotelis Kazakopoulos, Professor IHU
Dimitrios Stamovlasis, Associate Professor AUTH
Title:
Study of interfaces in microcontrollers electronic circuits learning
Summary:
The scope of this PhD thesis is to study the interfaces in microcontrollers electronic circuits learning. These interfaces can be categorized into Physical (Physical User Interfaces – PUIs) and Virtual (Virtual User Interfaces – VUIs). The main goal is to highlight the advantages, disadvantages and conditions under which one interface is more efficient than the other. More specifically, PUIs use real electronic components, while VUIs use computer simulations. In other words, the object of this research is the comparison of tangible electronic circuits with the corresponding simulations in software programs during the microcontrollers’ learning, but also the study of mixed PUI-VUI models. The key research question is which of the modes is more efficient and whether the sequence of interfaces affects the efficiency of a mixed model. Metric methods, such as the time needed to complete tasks, will be used to investigate the usability performance of physical and virtual interfaces.
PhD Candidate: Charalampos C. Charalampidis
Supervisor:
Panagiotis Adamidis, Professor IHU
Advisory Committee:
Athanasios Iosifidis, Associate Professor IHU
Efkleidis Keramopoulos Associate Professor IHU
Title:
Interoperability enhancement in the «Internet of Things», through utilization of «Semantic Web» technology
Summary:
The main objective of the present doctoral thesis, is the research for Semantic Web (SW) technology utilization in the logical representation of Internet of Things’ (IoT) elements, in order to enhance these elements’ interoperability. The motive for this thesis, is the interoperability deficit, observed among the various IoT elements. The main cause of this deficit is the plethora of protocols and technologies targeting IoT hardware and software, as well as the plethora of semantic models (vocabularies/ontologies), proposed for the logical representation of IoT elements. The interoperability deficit, complicates IoT applications creation without prior knowledge and expertise of the respective technologies, protocols or standards. The goal is to develop know-how in the utilization of Semantic Web technology, for the creation of practical IoT applications, along with IoT elements (“things” under control, IoT management software, IoT user interfaces, etc) featuring enchanced interoperability.